Jason M. Lemkin discusses whether software companies will shift from selling seats to selling outcomes as advanced reasoning models like Deep SEQ become cheaper and more capable. If models get better and cheaper, companies will use more of them to improve user experience while keeping margins stable.
How long, how, how often can you charge two bucks for, for what costs a fraction of a cent to deliver? Yeah, I, maybe, but like software has never been cost based pricing, right? I mean, Twilio. Okay, fine. They stack a margin on top of the SMS cost to the underlying. Yeah, stripe, stripe. Okay, but, but like a lot of APIs are cost based. But keep going. I'm, I'm, I'm with you, I'm with you. I mean like gainsight, you know, they're not marking up AWS storage. Right. Or like compute. So I hear what you're saying, but I do think the point that you made at the beginning I completely agree with, which is like if Deep SEQ like models with great reasoning, 95% cheaper software companies will just like 10x or 100x the use and probably keep the margins the same to get to a much better user experience as a way of differentiation. Right. So I think that the ux our expectations as users for what these things are going to do are going to go way up because they become increasingly sophisticated. And like if you play with Deep seq, I was trying to create a flowchart. I took the Deep SEQ paper, the academic paper, and I put it in a Deep SEQ and I said create a flowchart for me of what they did because we could read an academic paper. It's been two and a half minutes thinking and I was watching it at first it said, okay, like the user said create a single chart. But the experiments actually have three branches. So now I have to decide like, do I create one single chart or do I create a chart with three different branches? And it like went back and forth, back and forth, back and forth. And now I understand. Okay, well next time I ask it to make a chart, I need to be clear on this point. Just the same way that we learned how to use Google, we're learning how to use these models. And this is one of the first models that narrates the way that it thinks. And I think that way we'll start to get to more outcomes. And then I think the other question that's really relevant to this conversation is are we selling outcomes now in software as opposed to selling seats? Yeah. And I don't know the answer, but the mantra has been we're selling outcomes. We're selling outcomes. We're selling leads, not SDR seats. We're selling resolved customer support cases or videos or clicks.
